[ale] Hard Drive Failures
Neal Rhodes
neal at mnopltd.com
Fri Feb 22 17:04:10 EST 2013
All the comments have been informative, and while not directly
disagreeing with anyone, I will classify backup/restore into several
categories:
A. The stuff needed to make the box come back up on its hind legs and
dance. It is really useful to have a bare-iron bootable recover for
this. It doesn't change terribly often, only maybe monthly, so
hopefully I can remember to re-do the hardware image. The problem is,
after 9 months, I'm unlikely to want to restore that same OS image to
the same hardware because the dang world has changed again. If it's an
IDE drive, I'll want a SATA drive. If it's Centos 6, then I'll want it
to be Centos 7.
B. Applications which are a pain to re-install and re-configure. This
is obviously easier on Linux than that other OS.
C. My stuff which may exist nowhere else in the world. It makes sense
to do a file-based incremental backup daily and get this offsite. I
want this backup to be readable by any OS. If all the computers in the
office get stolen, I've got to be able to work off the Ubuntu/Vista
notebook.
D. Stuff which does exist elsewhere, like PDF manual sets, music I own,
stuff that is easy to install.
So, ideally, (IMHO) I should use a bare-metal bootable strategy on the
OS, (A) but an incremental file based strategy on C, and maybe not even
bother on D, and how to handle B just depends.
I try to keep a bright red line between the / filesystem versus /home
and always keep "my stuff" on /u, and I'm starting to try to distinguish
between /u "my stuff" and /u2 "stuff I don't care about".
Neal
More information about the Ale
mailing list